Early Growth Response Genes Increases Rapidly
After Mechanical Overloading and Unloading in
Tendon Constructs
Andreas Herchenhan, Franciele Dietrich, Peter Schjerling, Michael Kjaer and Pernilla T. Eliasson
The self-archived postprint version of this journal article is available at Linköping University Institutional Repository (DiVA):
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-162512
N.B.: When citing this work, cite the original publication.
Herchenhan, A., Dietrich, F., Schjerling, P., Kjaer, M., Eliasson, P. T., (2019), Early Growth Response Genes Increases Rapidly After Mechanical Overloading and Unloading in Tendon Constructs, Journal of Orthopaedic Research. https://doi.org/10.1002/jor.24513
Original publication available at:
https://doi.org/10.1002/jor.24513
Copyright: Wiley (12 months)
1 Early Growth Response genes increases rapidly after mechanical overloading and unloading in tendon constructs.
Andreas Herchenhan1; Franciele Dietrich-Zagonel2, Peter Schjerling1; Michael Kjær1;
Pernilla Eliasson1,2
Affiliations:
1Institute of Sports Medicine Copenhagen, Department of Orthopedic Surgery M, Bispebjerg
Hospital and Center for Healthy Aging, Faculty of Health and Medical Sciences, University of Copenhagen, Denmark.
2 Department of Clinical and Experimental Medicine, Division of Surgery, Orthopedics and
Oncology (KOO), Linköping University, Linköping, Sweden.
Address for corresponding author: Pernilla Eliasson
Orthopaedics, Cell biology building, level 10, Linköping University, 58185 Linköping, Sweden Department of Clinical and Experimental Medicine,
Telephone: +46 101034271
e-mail: pernilla.eliasson@gmail.com
2 Author contributions statement:
PE, AH, PS, and MK planned the study. PE, AH, and FDZ performed the experiments. All authors were involved in the data analysis. PE wrote the first manuscript draft. All authors revised and approved the final manuscript.
3 Abstract
Tendon cells exist in a dense extracellular matrix (ECM) and mechanical loading is important for the strength development of this matrix. We therefore use a three-dimensional (3D) culture system for tendon formation in vitro. The objectives of this study were to elucidate the temporal expression of tendon related genes during the formation of artificial tendons in vitro and to investigate if early growth response (EGR)-1, EGR-2, cFOS, cyclooxygenase (Cox)-1 and -2 are sensitive to mechanical loading. First, we studied mRNA levels of several tendon related genes during formation of tendon constructs. Secondly, we studied the mRNA levels of e.g. EGR-1 and EGR-2 after different degrees of loading; dynamic physiologic-range loading (2.5% strain), dynamic overloading (approximately 10% strain) or tension release. The gene expression for tendon related genes (i.e. EGR-2, Mohawk, tenomodulin, collagen type 3) increased with time after seeding into this 3D model. EGR-1, EGR-2, cFOS, Cox-1, and Cox-2 did not respond to physiologic-range loading. But overloading (and tension release) lead to elevated levels of EGR-1 and EGR-2 (p≤0.006). cFOS and Cox-2 were increased after overloading (both p<0.007) but not after tension release (p=0.06 and 0.08). In conclusion, the expression of tendon related genes increases during the formation of artificial tendons in vitro, including EGR-2. Furthermore, the gene expression of EGR-1 and EGR-2 in human tendon cells appear to be sensitive to overloading and unloading but did not respond to the single episode of physiologic-range loading. These findings could be helpful for the understanding of tendon tensional homeostasis.
4 Introduction
Mechanical forces are important for intact tendon maintenance as well as proper tendon repair 1; 2. Loading gives strength to the extracellular matrix and complete unloading during
tendon healing result in poor material properties 3. However, the exact mechanism behind the
positive effect of loading during tendon healing remains unclear. TGF-β/SMAD2/3 and FGF/ERK/MAPK are two pathways which are believed to be important for tendon
development as well as in the mechanotransduction process 4. However up-stream regulators
of these two pathways are not clear.
The transcription factors, early growth response (EGR)-1 and EGR-2 are two immediate early genes which can be rapidly upregulated in response to a stimulus. They are important for both tendon development and tendon repair where they drive the expression of scleraxis (SCX) and collagen type 1 5; 6, but they are not specific for tendons. The EGR genes have
been shown to be mechanosensitive in bone, muscle and endothelial cells 7-10. Also, tension
release on tendon cells have been shown to down-regulate the mRNA levels of EGR-1 11.
Treadmill exercise of rats with healing Achilles tendons have been shown to induce an increased expression of both EGR-1 and EGR-2 12. The tissue, during early tendon healing is
weak 1 as why unrestricted walking and full loading can induce micro damages and small
bleedings in the tissue 13; 14. Skin wounding in mice can induce an increased expression of
EGR-1 and EGR-2 which are believed to work as kick-start genes for the healing response 8.
Furthermore, needling induced microdamage in healing rat Achilles tendon can also lead to an increased gene expression of these two genes 15. It is therefore difficult to distinguish
between an effect of mechanotransduction and an effect of microdamage induced inflammation during tendon healing in a rat model.
5 In vitro studies can either be performed by two-dimensional (2D) culturing or
three-dimensional (3D) culturing, also in studies on the effect of loading. 3D culturing can be performed in many ways, e.g. scaffold free or be based on collagen or fibrin as the primary scaffold 16-20. In this study, we use a fibrin-based 3D culture system with primary human
tendon fibroblasts. Cells are seeded out in a fibrin gel which contracts between two anchor points to form a linear tendon like tissue, referred to as tendon construct. The cells produce collagen in a parallel manner and arrange them self in a spindle like shape between the
collagen fibrils and this gives a static tension to the constructs. The tendon constructs can also be loaded dynamical with uniaxial stretch on all the cells, something that’s not possible with conventional 2D cell culture system.
As in our in vitro system, tendon healing in vivo is initiated by a fibrin clot which acts as a preliminary scaffold for cells and matrix. This model can thereby be used as a model to study the effect of loading without microdamage induced inflammation, which is seen after loading during early tendon healing in rats. The objectives of this study were to elucidate the
temporal expression of tendon related genes in human tendon cells during the formation of a 3D matrix in vitro and to investigate whether EGR-1, EGR-2 and cFOS are sensitive to changes in loading. This was done in four separate experiments with different specific aims; 1) to elucidate how tendon related markers are expressed during tendon construct formation, 2) to study changes in gene expression after a short bout of dynamic loading within the physiologic-range, 3) to investigate the role of preconditioning before dynamic loading, and finally 4) to elucidate if extreme changes in loading alter the gene expression of EGR-1 and EGR-2.
6 Materials and methods
Isolation of cells
Tendon fibroblast were isolated from human semitendinosus and gracilis tendons, from patients undergoing reconstructive anterior cruciate ligament surgery. Muscle tissue was carefully removed, and tendon tissue was cut into small pieces. The tissue was digested overnight in DMEM/F12 (Gibco, Invitrogen) supplemented with 0.1% collagenase type II (Worthington) and 20% fetal bovine serum (FBS, Gibco, Invitrogen). Cells were seeded out in flasks the following day and cultured to confluence in DMEM/F12 supplemented with 10% FBS. The method has previously been described in detail 21. All cells were between the
1st to 4th passage and came from 18–32 years old patients. The present experiment was approved by local ethical committees (refs. H-3-2010-070; 2015/408-31) and all patients gave written informed consent.
Tendon construct formation
Tendon constructs were assembled as previously described 21. Six-well plates were coated at
the bottom with SYLGARD (Dow-Chemicals) and two short silk sutures (0.5 cm, Ethicon) were pinned onto this. The plates were sterilized by ethanol immersion and 250 000 cells were seeded in each well within a fibrin gel on top of the SYLGARD. The fibrin gel
consisted of human tendon fibroblasts in DMEM/F12 containing 10% FBS, 4 mg/mL human fibrinogen, 10 μg/mL aprotenin and 1 unit of human thrombin (all Sigma Aldrich). The fibrin gel was allowed to set for 60 minutes at 37°C before it was covered with DMEM/F12
(supplemented with 10% FBS, 0.2 mM L-ascorbic acid 2-phosphate, 0.05 mM L-Proline and 1% Penicillin-Streptomycin). The medium was replaced every second to third day and adhesions to the side of the well were detached using a fine pipette tip to allow gel
7 contraction. After 8–14 days the cells contracted the structure to a rod-like structure in
between the anchor points
Experimental setup
This study consisted of 4 separate experiments (Figure 1).
Experiment 1: Cells were seeded in the fibrin gel as described above and samples were harvested during the formation of tendon constructs at 1, 2, 4, 8, and 16 days after seeding. Real-time PCR was used to study the mRNA expression of several tendon related genes; EGR-1, EGR-2, tenomodulin, scleraxis, mohawk, decorin, fibromodulin collagen type 1, 3, 5, 12, and 14. Cells from 5 different patients was used for this experiment. One sample in the 16 days group was lost due to contamination.
Experiment 2: Constructs were transferred to a bioreactor 14 days after cell seeding and subjected to dynamic loading. The loading protocol was adapted from Paxton et al 23 and
consisted of 5 minutes of cyclic uniaxial mechanical stretch (sinusoidal strain, 2.5% strain at 1Hz). 5 minutes was also chosen based on a previous in vivo experiment 12. The bioreactor
used in this study was custom-made by David Holmes at the University of Manchester and allows for two constructs to be stretched at the same time by a single stepper motor 17; 22.
Loading was performed in pre-warmed medium (37 °C), but the bioreactor was kept at room-temperature (23 °C). Loaded constructs were pinned back in the SYLGARD after loading in the bioreactor. Controls were kept in the SYLGARD with static loading, at room temperature, while the other constructs were subjected to dynamic loading. All constructs were thereafter placed back in the cell culture incubator. Loaded samples were harvested for real-time PCR after 15 minutes, 1h, 2h, 4h, 8h, 12h, 24h, and 48h. Controls were harvested after 15 minutes, 12 h, 24h and 48h. Cells from 5 different patients were used.
8 Experiment 3: Dynamic loading (sinusoidal strain, 2.5% strain, 1Hz, 5 min) was performed either at room temperature immediate after transfer to the bioreactor (transferred stretched) or at 37°C after 24 hours of preconditioning (preconditioned stretched). Sham treated groups (transferred sham and preconditioned sham) were transferred to the bioreactor but was left for 5 minutes with static loading (no dynamic loading). Two statically loaded
controls were used, one was harvested immediately after the plate from 37°C when the plate was taken out of the incubator (untouched control), the other control was kept at room temperature (sham control) while other constructs were dynamically loaded. All loaded and sham treated samples were repositioned in the SYLGARD after stretching and harvested one hour later. Cells from 5 different patients were used for this experiment and the constructs were used 14 days after cell seeding.
Experiment 4: Tendon constructs, 14 days after seeding, were either released completely from tension (tension release) or exposed to 5x10% stretch (pull 10%). This stretch was performed manually (no bioreactor). The anchor suture was unpinned from the SYLGARD on one side of the construct and pulled 1.5 mm by forceps, giving an approximately strain of 10%. Each stretching cycle took approximately 3 seconds, 1 second to reach 1.5 mm, 1 second in this position, and 1 second to return to the original position.This was repeated five times before the suture was pinned back in the SYLGARD. 10% strain is considered to be above the stretch level where micro-ruptures begins to occur 24. Tension release was
performed by releasing the anchor on one side. Samples were harvested 15 minutes later and cells from 6 different patients were used. Two control groups were used as in experiment 3 (untouched control and sham control).
9 Quantitative real-time PCR
The mRNA expression of different target genes was measured using quantitative real-time reverse transcriptase (RT) PCR. Targets and primer sequences are provided in Table 1. Tendon constructs were quickly rinsed in PBS before harvesting and thereafter transferred to tubes containing 1 mL TriReagent (Molecular Research Centre, Cincinnati, OH, USA), 5 stainless steel beads (2.3 mm in diameter) and 5 silicon-carbide sharp particles (1 mm in diameter) for mechanical disruption (BioSpec Products, Inc., Bartlesville, Oklahoma, USA). Samples collected at day 1, 2 and 4 in experiment 1 were divided into smaller pieces with a scalpel and smaller parts of the gel was used for RNA extraction to prevent too high
concentrations of water in relation to the TriReagent level. Samples were mechanically disrupted using a FastPrep®-24 instrument (MP Biomedicals, Inc., Illkirch, France), and subsequently 100 µL bromo-chloropropane (Molecular Research Centre) was added in order to separate the samples into an aqueous and an organic phase.120 µg/mL of glycogen was added to each sample to improve RNA yield. Following isolation of the aqueous phase, RNA was precipitated using isopropanol, washed in ethanol and dissolved in RNase-free water. RNA concentrations and purity were determined by spectroscopy at 260, 280, 240 and 320 nm and RNA quality was confirmed by gel electrophoresis.
The yield of RNA was 2.3±3.1µg/sample in experiment 1, 4.1±0.8 µg/construct in
experiment 2, 3.8±1.2 µg in experiment 3 and 4.3±1.2 µg in experiment 4. We used 330 ng of total RNA for cDNA synthesis in experiment 1, 850 ng in experiment 2, 850 ng in
experiment 3, and 500 ng in experiment 4. RNA was transcribed to complementary DNA (cDNA) by using Omniscript reverse transcriptase (Qiagen, Hilden, Germany). For each target gene, 5 μL of 20x diluted cDNA (in 10mM Tris, 1 mM EDTA buffer, pH 8 with 1 ng/μL salmon DNA) was amplified in 25 μL Quantitect SYBR Green Master Mix (Qiagen) with specific primers (100 nM each) on a real-time PCR machine (MX3005P, Stratagene, La
10 Jolla, CA, USA). The thermal profile was 95 °C, 10 min → (95 °C, 15 s → 58 °C, 30 s → 63 °C, 90 s) × 50 → 95 °C, 60 s → 55 °C, 30 s → 95 °C, 60 s. Signal intensity was acquired at the 63 °C step, and the threshold cycle (Ct) values were related to a standard curve made with the cloned PCR product. Specificity was confirmed by melting curve analysis after
amplification (the 55 °C to 95 °C step). The large ribosomal protein P0 (RPLP0) mRNA was chosen as internal control for normalization. To test RPLP0 mRNA stability, another
common mRNA control was also measured, Glyceraldehyde 3-phosphate dehydrogenase (GAPDH). The GAPDH/RPLP0 ratio appeared to be stable (no significant difference between the groups) and RPLP0 was chosen for normalization.
Statistical analysis
Experiment 1: all samples were normalized to day 1 for each patient. The effect of time from seeding (day 1 to day 16) was analyzed with a Kruskall-Wallis test, followed by Dunn´s post hoc test (comparing each time-point to day 1).
Experiment 2: The four control groups (C, C12, C24, C48) were first analyzed with a Friedmans test. If no difference was found; samples were normalized to C for each patient and analyzed with a Kruskall-Wallis test, followed by Dunn´s post hoc test (comparing each time-point to the control).
Experiment 3: all samples were normalized to the untouched control for each patient and analyzed with a Kruskall-Wallis test, followed by Dunn´s post hoc test (comparing:
untouched control vs normal control; untouched control vs sham stretched; sham stretched vs stretched; and preconditioned samples sham stretched vs stretched).
Experiment 4: was first analyzed by comparing the controls (untouched and sham) with a Wilcoxon matched-pairs signed rank test. If no difference was found; samples were
11 normalized to the untouched control for each patient and analyzed with a Kruskall-Wallis test, followed by Dunn´s post hoc test.
12 Results
Experiment 1: the gene expression of several tendon related genes increases with time after seeding in 3D culture model.
mRNA levels of EGR-2, Mohawk, collagen type 3, 12 and 14, decorin and fibromodulin were all influenced by time after seeding (Figure 2). EGR-2 was approximately 11 times higher at day 16, when the construct is formed, compared to day 1 (p=0.04). Similarly, to EGR-2, EGR-1 slowly increased with time although not statistically significant (p=0.07). Mohawk was 5 times more expressed at day 16 compared to day 1 (p=0.006) while scleraxis was unaffected by time (p=0.09). Tenomodulin was only measurable in one sample day 1 in contrast to all samples at day 16. Collagen type 14 had the highest increase with time, with 19 times higher expression at day 16 compared to day 1 (p=0.005). The expression of collagen type 3 peaked at day 8 (p=0.006) and collagen type 12 was approximately 3 times more expressed at day 16 (p=0.002). Collagen type 1 and 5 were not significantly different over time (p=0.5 and p=0.6 respectively). The expression of decorin and fibromodulin was both increased with time from day 1 to day 16 (p=0.009 and p=0.002 respectively).
Experiment 2: The expression of EGR-1 and cFOS peaks at 1 hour after a stretching episode. There was no difference between the control groups (C, C12, C24 or C48) for any of the genes. EGR-1 and cFOS were significantly increased in the stretched group (p<0.05 and p<0.001 respectively) with a peak at 1 hour (Figure 3). Both genes were back to baseline levels 8 hours after stretching. EGR-2 tended to be increased with a peak at 2 hours, although this was not statistically significant (p=0.06). Cox-1 and Cox-2 had also a transient increased expression (p<0.01 and p<0.0001 respectively). Cox-2 peaked at 4 hours, and was then 25 times more expressed than control. Cox-1 increased more slowly and peaked at 24 hours. Both collagen type 1 and 3 were unaltered (p=0.4 and p=0.2 respectively).
13 Experiment 3: EGR-1 and EGR-2 are increased after transfer to the bioreactor but not by physiologic-range loading.
EGR-1, EGR-2, cFOS, Cox-1, and Cox-2 were all significantly altered in the Kruskall-Wallis test (p<0.001 for all). But post hoc analysis showed an increased expression in the immediate transferred groups (both the sham and stretched) compared to control (Figure 4). There was no difference between the sham and stretched groups (irrespective of preconditioning or not).
Experiment 4: Extreme changes in load alters the expression of EGR-1 and EGR-2. There was no difference between the control groups (sham and untouched) for any gene. EGR-1, EGR-2, cFOS, and Cox-2 were all altered 15 minutes after either tension release or pulling to approximately 10% strain (p≤0.006 for all, Figure 5). The gene expression for EGR-1 was elevated by both increased and decreased loading, but there was a more pronounced response after increased loading (p=0.04 for tension release and 0.001 for pulling). The same pattern was seen for EGR-2 (p=0.01 and p=0.002). cFOS and Cox-2 was increased after pulling (both p<0.007) but not after tension release (p=0.06 for cFOS and 0.08 for Cox-2).
14 Discussion
The expression of tendon related genes - EGR-2, tenomodulin, Mohawk, collagen type 3, type 12, type 14, decorin, and fibromodulin were all increased with time during construct formation. Smaller collagen molecules, e.g. type 3, 12, and 14 were all more expressed in the formed constructs compared to day 1. However, the main tendon collagen, type 1, was not increased with time. Handling of the constructs (transfer and mounting in the bioreactor) lead to an increased expression of EGR-1, EGR-2, cFOS, and Cox-2. These genes were also increased by manual overloading (approximately 10% strain), but not by smaller magnitude (2.5%) dynamic stretching in the bioreactor. The increase in EGR-1 and -2 was rapid and peaked after 1-2 hours. This indicates that the early growth response genes have a rapid response to overload or unloading but not to a single bout of physiologic-range tensile loading in tendon cells.
EGR-2 increased to a larger extent than EGR-1 during constructs formation, resulting in a higher expression within the established constructs. EGR-1 and EGR-2 have previously been shown to be important for tendon formation and tendon development, but with different localization 6. EGR-1 is primarily expressed close to the myotendinous junction or in the
surrounding of the longitudinal tendons while EGR-2 is expressed throughout the tendon. Our model with tendon constructs probably represents more the general tendon cells and not specifically those of the myotendinous junction. This can therefore explain the higher
expression of EGR-2 compared to EGR-1. The most pronounced expression of EGR-2 was at 8 and 16 days after seeding, when the construct is formed. Day 16 was also the single time-point when tenomodulin reached detectable levels in all samples. Another important tendon related gene, Mohawk, was also increased with time during the formation of the constructs.
15 This indicates that our model is somewhat more similar to the in vivo situation, even if it is still far from a human tendon.
EGR-1 and -2 have previously been shown to respond to 5 minutes of treadmill exercise in a rat model for Achilles tendon healing 12. The second experiment was performed to study if
these genes responded to a short bout of dynamic loading in tendon constructs. The stretching protocol used in this experiment has previously been shown to induce a ERK phosphorylation and lead to an increase in collagen levels 23. EKR is regarded a key point upstream to
mechanical signals and cFOS is depended on ERK1/2 activity 25. Furthermore, EGR-2 has
been shown to be ERK1/2 mediated after loading in bone cells 8. Our experiment showed a
transient effect on most studied genes (except collagen type 1 and 3). However, the use of our bioreactor included several steps; transfer of the construct, mounting, stretching and finally a transfer back to the original plate. The bioreactor was also kept at room temperature. These are all potential sources of errors. The third experiment was performed to control for several factors, both handling of the construct before the stretch and the temperature. This showed that transfer and mounting of the constructs lead to increased levels of EGR-1 and -2 but not the physiological-range loading. The transfer can potentially lead to short sessions of either unloading or more “heavy” loading during mounting. The last study was performed to
investigate if unloading or acute overloading leads to changes in the expression of EGR-1 and -2. The results confirmed that both these changes can induce a change in the expression.
Strong loading of healing Achilles tendons lead to a higher gene expression of EGR-1 compared to both unloaded and mildly loaded animals (unpublished data). Furthermore, induced tissue damage by needling can also lead to increased levels of EGR-1 and EGR-2 15.
16 Thereby, the increased expression of EGR-1 and EGR-2 after loading in rats might be due to induced microtrauma and small tissue damage 14 and not due to mechanotransduction.
EGR-1 has been shown to be mechanosensitive in other cell types. Vascular smooth muscle cells, strained at 1Hz, have increased mRNA levels after one single cycle 7. Although, these
cells were exposed to 25% strain in a 2D model, and there was a threshold at 10% strain with one single loading cycle to induce increased mRNA levels of EGR-1. With increased time with 30 minutes of loading, there was instead only a need for 5% strain to induce a response. EGR-1 has previously been pointed out to be mechanosensitive in tendon constructs made from mouse mesenchymal cells, and tension release lead to reduced expression for EGR-1, Scx, Tnmd, Col1a1, Col1a2 and TGFb2 11. This contrasts with our results. However, our
experiments are studies of the acute response (after 1 hour) while the previous study measured the expression after 24 hours. It appears as the gene expression of EGR-1 and EGR-2 are sensitive to acute overloading or unloading but not to physiological loading by 2.5% strain in tendon constructs. At least not with a short time stretching at 2.5% strain. It is possible that these genes mainly respond to “injury” and not to non-destructive stretch. The perspective of these findings could be important for the understanding of tendon overloading and development of clinical tendinopathy. There are some limitations in this study. There was a difference in specimen lengths between experiment 2+3 and experiment 4 (10mm vs. 15mm), this could possibly have an impact on construct geometry and cell to matrix ratio. This might have some influence on the cellular response. The difference in temperature during loading (room temperature vs cell culture incubator) could potentially influence the gene expression. However, experiment 3 and 4 used two controls, one kept in the cell culture incubator and one in room temperature and this did not lead to and significant changes between the groups. Finally, the estimation of the overload (10% strain) is an approximate
17 calculation, while the 2.5% is an exact measurement. Albeit, this experiment was aiming to investigate if handling of the construct and possibly overloading them might have been the source for gene expression changes, and the results is in line with this. The fact that we unintendedly introduced gene expression changes during handling and transfer to the
bioreactor shows the importance of sham treated samples or preconditioning before studying the response to stretch.
In conclusion, 3D culturing of human tendon cells using a fibrin-based model leads to a time-dependent increase in the gene expression of tendon related genes, including EGR-2.
Furthermore, the gene expression of EGR-1 and EGR-2 in tendon constructs made by human tendon cells appear to be sensitive to overloading or unloading but not to a single bout of physiologic-range dynamic loading.
Acknowledgement
This study was supported by contract grant sponsors: Lundbeck foundation; Nordea foundation (Center of Healthy Aging); IOC Sports medicine Copenhagen; Danish Medical Research Council; Swedish Society for Medical Research; Lions research foundation, Magnus Bergvall foundation; Swedish Research Council for Sport Science; and Swedish fund for research without animal experiments.
18 References
1. Eliasson P, Andersson T, Aspenberg P. 2009. Rat Achilles tendon healing: mechanical loading and gene expression. J Appl Physiol (1985) 107:399-407. 2. Magnusson SP, Kjaer M. 2018. The impact of loading, unloading, ageing and injury
on the human tendon. J Physiol.
3. Andersson T, Eliasson P, Hammerman M, et al. 2012. Low-level mechanical stimulation is sufficient to improve tendon healing in rats. J Appl Physiol (1985) 113:1398-1402.
4. Havis E, Bonnin MA, Esteves de Lima J, et al. 2016. TGFbeta and FGF promote tendon progenitor fate and act downstream of muscle contraction to regulate tendon differentiation during chick limb development. Development 143:3839-3851.
5. Guerquin MJ, Charvet B, Nourissat G, et al. 2013. Transcription factor EGR1 directs tendon differentiation and promotes tendon repair. J Clin Invest 123:3564-3576. 6. Lejard V, Blais F, Guerquin MJ, et al. 2011. EGR1 and EGR2 involvement in
vertebrate tendon differentiation. J Biol Chem 286:5855-5867.
7. Morawietz H, Ma YH, Vives F, et al. 1999. Rapid induction and translocation of Egr-1 in response to mechanical strain in vascular smooth muscle cells. Circ Res 84:678-687.
8. Zaman G, Sunters A, Galea GL, et al. 2012. Loading-related regulation of
transcription factor EGR2/Krox-20 in bone cells is ERK1/2 protein-mediated and prostaglandin, Wnt signaling pathway-, and insulin-like growth factor-I axis-dependent. J Biol Chem 287:3946-3962.
9. Pardo PS, Mohamed JS, Lopez MA, et al. 2011. Induction of Sirt1 by mechanical stretch of skeletal muscle through the early response factor EGR1 triggers an antioxidative response. J Biol Chem 286:2559-2566.
19 10. Stula M, Orzechowski HD, Gschwend S, et al. 2000. Influence of sustained
mechanical stress on Egr-1 mRNA expression in cultured human endothelial cells. Mol Cell Biochem 210:101-108.
11. Gaut L, Robert N, Delalande A, et al. 2016. EGR1 Regulates Transcription Downstream of Mechanical Signals during Tendon Formation and Healing. PLoS One 11:e0166237.
12. Eliasson P, Andersson T, Hammerman M, et al. 2013. Primary gene response to mechanical loading in healing rat Achilles tendons. J Appl Physiol (1985) 114:1519-1526.
13. Eliasson P, Andersson T, Aspenberg P. 2012. Achilles tendon healing in rats is
improved by intermittent mechanical loading during the inflammatory phase. J Orthop Res 30:274-279.
14. Hammerman M, Dietrich-Zagonel F, Blomgran P, et al. 2018. Different mechanisms activated by mild versus strong loading in rat Achilles tendon healing. PLoS One 13:e0201211.
15. Hammerman M, Aspenberg P, Eliasson P. 2014. Microtrauma stimulates rat Achilles tendon healing via an early gene expression pattern similar to mechanical loading. J Appl Physiol (1985) 116:54-60.
16. Garvin J, Qi J, Maloney M, et al. 2003. Novel system for engineering bioartificial tendons and application of mechanical load. Tissue Eng 9:967-979.
17. Giannopoulos A, Svensson RB, Heinemeier KM, et al. 2018. Cellular homeostatic tension and force transmission measured in human engineered tendon. J Biomech 78:161-165.
20 18. Kalson NS, Holmes DF, Kapacee Z, et al. 2010. An experimental model for studying
the biomechanics of embryonic tendon: Evidence that the development of mechanical properties depends on the actinomyosin machinery. Matrix Biol 29:678-689.
19. Kuo CK, Tuan RS. 2008. Mechanoactive tenogenic differentiation of human mesenchymal stem cells. Tissue Eng Part A 14:1615-1627.
20. Larkin LM, Calve S, Kostrominova TY, et al. 2006. Structure and functional evaluation of tendon-skeletal muscle constructs engineered in vitro. Tissue Eng 12:3149-3158.
21. Bayer ML, Yeung CY, Kadler KE, et al. 2010. The initiation of embryonic-like collagen fibrillogenesis by adult human tendon fibroblasts when cultured under tension. Biomaterials 31:4889-4897.
22. Holmes DF, Yeung CC, Garva R, et al. 2018. Synchronized mechanical oscillations at the cell-matrix interface in the formation of tensile tissue. Proc Natl Acad Sci U S A 115:E9288-E9297.
23. Paxton JZ, Hagerty P, Andrick JJ, et al. 2012. Optimizing an intermittent stretch paradigm using ERK1/2 phosphorylation results in increased collagen synthesis in engineered ligaments. Tissue Eng Part A 18:277-284.
24. Snedeker JG, Foolen J. 2017. Tendon injury and repair - A perspective on the basic mechanisms of tendon disease and future clinical therapy. Acta Biomater 63:18-36. 25. Liu J, Liu T, Zheng Y, et al. 2006. Early responses of osteoblast-like cells to different
mechanical signals through various signaling pathways. Biochem Biophys Res Commun 348:1167-1173.
21 Figure legends
Figure 1. Experimental setup.
This study was divided into 4 separate experiments. Experiment 1: samples were harvested at 1, 2, 4, 8, and 16 days after seeding. Experiment 2: Samples were harvested after 15 minutes, 1h, 2h, 4h, 8h, 12h, 24h, and 48h after 5 minutes stretch at 2.5% strain at 1Hz. Controls (no stretch) were harvested at 15 minutes, 12 h, 24h and 48h. Experiment 3: two controls were used in this experiment, one which was harvested immediately after the plate had been taken out of the incubator (untouched control) and one that was left at room temperature (sham control). Stretched constructs were subjected to 5 minutes of stretch at 2.5% strain and 1Hz as in experiment 2, either immediate after transfer to the bioreactor (transferred stretched) or after preconditioning (preconditioned stretched). Additionally, sham treated constructs were used which were unpinned, transferred to the bioreactor but was left for 5 minutes at room temperature, no stretch (transferred sham) or they were preconditioned and were left as sham with no stretch (preconditioned sham). All samples were harvested 1 hour later. Experiment 4: tendon constructs were either released completely from tension or exposed to 5x10% stretch (pull 10%). The samples were harvested 15 minutes later.
Figure 2. Gene expression analysis: experiment 1.
Normalized gene expression at 1, 2, 4, 8 and 16 days after seeding. Day 1 is set as baseline, and mRNA expression at the other time-points are shown relative to day 1 for each cell line. Data is presented on a logarithmic y scale. Individual donor-specific cell lines are represented by the different symbols. n =5 for day 1, 2, 4 and 8 and n=4 for day 16. Significant changes from d1 are indicated by * (p<0.05).
22 Figure 3. Gene expression analysis: experiment 2.
Normalized gene expression at 15 minutes, 1h, 2h, 4h, 8h, 12h, 24h, and 48h after 5 minutes dynamic stretch at 2.5% strain at 1Hz. Control is set as baseline and mRNA expression in the other groups are shown relative to the control for each cell line. Data is presented on a
logarithmic y scale. Individual donor-specific cell lines are represented by the different symbols. n =5. Significant changes from controls are indicated by * (p<0.05).
Figure 4. Gene expression analysis: experiment 3.
Normalized gene expression after stretching or sham treatment in constructs that have been preconditioned or not. Untouched control is set as baseline and mRNA expression in the other groups are shown relative to this for each cell line. Data is presented on a logarithmic y scale. Individual donor-specific cell lines are represented by the different symbols. n =5. Significant changes from controls are indicated by * (p<0.05).
Figure 5. Gene expression analysis: experiment 4.
Normalized gene expression after tension release or pulling to 10% strain. Control is set as baseline and mRNA expression in the other groups are shown relative to this for each cell line. Data is presented on a logarithmic y scale. Individual donor-specific cell lines are represented by the different symbols. n =6. Significant changes from controls are indicated by * (p<0.05).
Table 1. Primers for real-time PCR.
mRNA Genbank Sense Antisense
RPLP0 NM_053275.3 GGAAACTCTGCATTCTCGCTTCCT CCAGGACTCGTTTGTACCCGTTG
GAPDH NM_002046.4 CCTCCTGCACCACCAACTGCTT GAGGGGCCATCCACAGTCTTCT
COL1A1 NM_000088.3 GGCAACAGCCGCTTCACCTAC GCGGGAGGACTTGGTGGTTTT
COL3A1 NM_000090.3 CACGGAAACACTGGTGGACAGATT ATGCCAGCTGCACATCAAGGAC
COL5A1 NM_000093.4 AGCAGATGAAACGGCCCCTG TCCTTGGTTAGGATCGACCCAGT
COL12A1 NM_080645.2 CCCAGGTCCTCCTGGATACTGTGA GCAGCACTGGCGACTTAGAAAATGT
COL14A1 NM_021110.3 AGCATGGGACCGCAAGGC GACGCGCCACTGATCTCACC
EGR-1 NM_001964.3 GCAGCCCTACGAGCACCTGACC AACTGGTCTCCACCAGCACCTTC
EGR-2 NM_000399.5 CCTTTGACCAGATGAACGGAGTG TAGGTGCAGAGACGGGAGCAAA
Scleraxis NM_001080514.3 CAGCCCAAACAGATCTGCACCTT CTGTCTTTCTGTCGCGGTCCTT
Mohawk NM_173576.3 CATCGTCATCAGAAACTGAAGGCA TCTGTTAGCTGCGCTTTCACCC
Decorin NM_001920.5 GGTGGGCTGGCAGAGCATAAGT TGTCCAGGTGGGCAGAAGTCA
Fibromodulin NM_002023.5 CAGTCAACACCAACCTGGAGAACC TGCAGAAGCTGCTGATGGAGAA
cFOS NM_005252.4 GGCCGGGGATAGCCTCTCTT GGCACTGGAGACGGCCAGGT
Cox-1 NM_000962.3 GGTTTGGCATGAAACCCTACACCT CCTCCAACTCTGCTGCCATCT
